The potential mechanism and clinical application value of remote ischemic conditioning in stroke

Some studies have confirmed the neuroprotective effect of remote ischemic conditioning against stroke. Although numerous animal researches have shown that the neuroprotective effect of remote ischemic conditioning may be related to neuroinflammation, cellular immunity, apoptosis, and autophagy, the exact underlying molecular mechanisms are unclear. This review summarizes the current status of different types of remote ischemic conditioning methods in animal and clinical studies and analyzes their commonalities and differences in neuroprotective mechanisms and signaling pathways. Remote ischemic conditioning has emerged as a potential therapeutic approach for improving stroke-induced brain injury owing to its simplicity, non-invasiveness, safety, and patient tolerability. Different forms of remote ischemic conditioning exhibit distinct intervention patterns, timing, and application range. Mechanistically, remote ischemic conditioning can exert neuroprotective effects by activating the Notch1/phosphatidylinositol 3-kinase/Akt signaling pathway, improving cerebral perfusion, suppressing neuroinflammation, inhibiting cell apoptosis, activating autophagy, and promoting neural regeneration. While remote ischemic conditioning has shown potential in improving stroke outcomes, its full clinical translation has not yet been achieved.

Dynamic Multi-Graph Spatio-Temporal Graph Traffic Flow Prediction in Bangkok:An Application of a Continuous Convolutional Neural Network

The ability to accurately predict urban traffic flows is crucial for optimising city operations.Consequently,various methods for forecasting urban traffic have been developed,focusing on analysing historical data to understand complex mobility patterns.Deep learning techniques,such as graph neural networks(GNNs),are popular for their ability to capture spatio-temporal dependencies.However,these models often become overly complex due to the large number of hyper-parameters involved.In this study,we introduce Dynamic Multi-Graph Spatial-Temporal Graph Neural Ordinary Differential Equation Networks(DMST-GNODE),a framework based on ordinary differential equations(ODEs)that autonomously discovers effective spatial-temporal graph neural network(STGNN)architectures for traffic prediction tasks.The comparative analysis of DMST-GNODE and baseline models indicates that DMST-GNODE model demonstrates superior performance across multiple datasets,consistently achieving the lowest Root Mean Square Error(RMSE)and Mean Absolute Error(MAE)values,alongside the highest accuracy.On the BKK(Bangkok)dataset,it outperformed other models with an RMSE of 3.3165 and an accuracy of 0.9367 for a 20-min interval,maintaining this trend across 40 and 60 min.Similarly,on the PeMS08 dataset,DMST-GNODE achieved the best performance with an RMSE of 19.4863 and an accuracy of 0.9377 at 20 min,demonstrating its effectiveness over longer periods.The Los_Loop dataset results further emphasise this model’s advantage,with an RMSE of 3.3422 and an accuracy of 0.7643 at 20 min,consistently maintaining superiority across all time intervals.These numerical highlights indicate that DMST-GNODE not only outperforms baseline models but also achieves higher accuracy and lower errors across different time intervals and datasets.

Actuator fault diagnosis and severity identification of turbofan engines for steady-state and dynamic conditions

Actuator faults can be critical in turbofan engines as they can lead to stall,surge,loss of thrust and failure of speed control.Thus,fault diagnosis of gas turbine actuators has attracted considerable attention,from both academia and industry.However,the extensive literature that exists on this topic does not address identifying the severity of actuator faults and focuses mainly on actuator fault detection and isolation.In addition,previous studies of actuator fault identification have not dealt with multiple concurrent faults in real time,especially when these are accompanied by sudden failures under dynamic conditions.This study develops component-level models for fault identification in four typical actuators used in high-bypass ratio turbofan engines under both dynamic and steady-state conditions and these are then integrated with the engine performance model developed by the authors.The research results reported here present a novel method of quantifying actuator faults using dynamic effect compensation.The maximum error for each actuator is less than0.06%and 0.07%,with average computational time of less than 0.0058 s and 0.0086 s for steady-state and transient cases,respectively.These results confirm that the proposed method can accurately and efficiently identify concurrent actuator fault for an engine operating under either transient or steady-state conditions,even in the case of a sudden malfunction.The research results emonstrate the potential benefit to emergency response capabilities by introducing this method of monitoring the health of aero engines.

FedCLCC:A personalized federated learning algorithm for edge cloud collaboration based on contrastive learning and conditional computing

Federated learning(FL)is a distributed machine learning paradigm for edge cloud computing.FL can facilitate data-driven decision-making in tactical scenarios,effectively addressing both data volume and infrastructure challenges in edge environments.However,the diversity of clients in edge cloud computing presents significant challenges for FL.Personalized federated learning(pFL)received considerable attention in recent years.One example of pFL involves exploiting the global and local information in the local model.Current pFL algorithms experience limitations such as slow convergence speed,catastrophic forgetting,and poor performance in complex tasks,which still have significant shortcomings compared to the centralized learning.To achieve high pFL performance,we propose FedCLCC:Federated Contrastive Learning and Conditional Computing.The core of FedCLCC is the use of contrastive learning and conditional computing.Contrastive learning determines the feature representation similarity to adjust the local model.Conditional computing separates the global and local information and feeds it to their corresponding heads for global and local handling.Our comprehensive experiments demonstrate that FedCLCC outperforms other state-of-the-art FL algorithms.

Riemann-Hilbert Problem and Multiple High-order Poles Solutions of the Focusing mKdV Equation with Nonzero Boundary Conditions

The focusing modified Korteweg-de Vries(mKdV)equation with multiple high-order poles under the nonzero boundary conditions is first investigated via developing a Riemann-Hilbert(RH)approach.We begin with the asymptotic property,symmetry and analyticity of the Jost solutions,and successfully construct the RH problem of the focusing mKdV equation.We solve the RH problem when 1/S_(11)(k)has a single highorder pole and multiple high-order poles.Furthermore,we derive the soliton solutions of the focusing mKdV equation which corresponding with a single high-order pole and multiple high-order poles,respectively.Finally,the dynamics of one-and two-soliton solutions are graphically discussed.

Successional conditions of Dolichospermum and Microcystis in Taihu Lake,China

Dolichospermum spp.and Microcystis spp.are two common cyanobacteria that form blooms in the Changjiang(Yangtze)River basin,but the environmental conditions for their succession in large lakes are still unclear.Based on daily monitoring data from Meiliang Bay in Taihu Lake from March to June,2016-2018,we studied the environmental conditions necessary for the succession of these two cyanobacteria.Results show that from March to June,the dominant genera of cyanobacteria experienced succession and co-dominated with Microcystis.The succession process included three stages.In StageⅠ,the biomass of Dolichospermum and Microcystis was similar(March),but Dolichospermum was dominant for most of the period.In StageⅡ,dominance alternated between Dolichospermum and Microcystis(April to mid-May).In StageⅢ,the biomass of Microcystis dominated(mid-May to June).In addition,temperature and nutrients across the three stages varied significantly.The average temperature increased continuously from 10.9 to 18.4,and to 24.2℃.The total nitrogen content decreased from 2.87 to 2.40,and to 1.86 mg/L.The total phosphorus content increased from 0.08 to 0.09,and to 0.12 mg/L.Correlation analysis revealed that Microcystis biomass was positively correlated with temperature and total phosphorus.Dolichospermum biomass was positively correlated with total nitrogen.Classification and regression tree displays that when the temperature was below 18.1℃,Dolichospermum dominated;above 18.1℃,Microcystis took over.Further analysis revealed that when temperature reached 18℃,the biomass of Microcystis increased exponentially,and the biomass of Dolichospermum exhibited a Gaussian distribution trend.This finding indicated that temperature was the key factor in the succession of Dolichospermum and Microcystis in nutrient-rich shallow lakes.As nitrogen and phosphorus concentrations decrease,the dominant species of cyanobacteria will diversify its development.The results of this study provide a foundation for risk prediction and control strategies for cyanobacterial blooms in lakes and reservoirs.

Conditioned medium from human dental pulp stem cells treats spinal cord injury by inhibiting microglial pyroptosis

Human dental pulp stem cell transplantation has been shown to be an effective therapeutic strategy for spinal cord injury.However,whether the human dental pulp stem cell secretome can contribute to functional recovery after spinal cord injury remains unclear.In the present study,we established a rat model of spinal cord injury based on impact injury from a dropped weight and then intraperitoneally injected the rats with conditioned medium from human dental pulp stem cells.We found that the conditioned medium effectively promoted the recovery of sensory and motor functions in rats with spinal cord injury,decreased expression of the microglial pyroptosis markers NLRP3,GSDMD,caspase-1,and interleukin-1β,promoted axonal and myelin regeneration,and inhibited the formation of glial scars.In addition,in a lipopolysaccharide-induced BV2 microglia model,conditioned medium from human dental pulp stem cells protected cells from pyroptosis by inhibiting the NLRP3/caspase-1/interleukin-1βpathway.These results indicate that conditioned medium from human dental pulp stem cells can reduce microglial pyroptosis by inhibiting the NLRP3/caspase-1/interleukin-1βpathway,thereby promoting the recovery of neurological function after spinal cord injury.Therefore,conditioned medium from human dental pulp stem cells may become an alternative therapy for spinal cord injury.

Uncertainties of landslide susceptibility prediction: Influences of random errors in landslide conditioning factors and errors reduction by low pass filter method

In the existing landslide susceptibility prediction(LSP)models,the influences of random errors in landslide conditioning factors on LSP are not considered,instead the original conditioning factors are directly taken as the model inputs,which brings uncertainties to LSP results.This study aims to reveal the influence rules of the different proportional random errors in conditioning factors on the LSP un-certainties,and further explore a method which can effectively reduce the random errors in conditioning factors.The original conditioning factors are firstly used to construct original factors-based LSP models,and then different random errors of 5%,10%,15% and 20%are added to these original factors for con-structing relevant errors-based LSP models.Secondly,low-pass filter-based LSP models are constructed by eliminating the random errors using low-pass filter method.Thirdly,the Ruijin County of China with 370 landslides and 16 conditioning factors are used as study case.Three typical machine learning models,i.e.multilayer perceptron(MLP),support vector machine(SVM)and random forest(RF),are selected as LSP models.Finally,the LSP uncertainties are discussed and results show that:(1)The low-pass filter can effectively reduce the random errors in conditioning factors to decrease the LSP uncertainties.(2)With the proportions of random errors increasing from 5%to 20%,the LSP uncertainty increases continuously.(3)The original factors-based models are feasible for LSP in the absence of more accurate conditioning factors.(4)The influence degrees of two uncertainty issues,machine learning models and different proportions of random errors,on the LSP modeling are large and basically the same.(5)The Shapley values effectively explain the internal mechanism of machine learning model predicting landslide sus-ceptibility.In conclusion,greater proportion of random errors in conditioning factors results in higher LSP uncertainty,and low-pass filter can effectively reduce these random errors.

Mixed D-vine copula-based conditional quantile model for stochastic monthly streamflow simulation

Copula functions have been widely used in stochastic simulation and prediction of streamflow.However,existing models are usually limited to single two-dimensional or three-dimensional copulas with the same bivariate block for all months.To address this limitation,this study developed a mixed D-vine copula-based conditional quantile model that can capture temporal correlations.This model can generate streamflow by selecting different historical streamflow variables as the conditions for different months and by exploiting the conditional quantile functions of streamflows in different months with mixed D-vine copulas.The up-to-down sequential method,which couples the maximum weight approach with the Akaike information criteria and the maximum likelihood approach,was used to determine the structures of multivariate Dvine copulas.The developed model was used in a case study to synthesize the monthly streamflow at the Tangnaihai hydrological station,the inflow control station of the Longyangxia Reservoir in the Yellow River Basin.The results showed that the developed model outperformed the commonly used bivariate copula model in terms of the performance in simulating the seasonality and interannual variability of streamflow.This model provides useful information for water-related natural hazard risk assessment and integrated water resources management and utilization.

Genesis of LCT Pegmatites during Early Paleozoic Orogeny of the North Qinling Orogenic Belt,China:Emplacement Conditions and Structural Control

The Guanpo pegmatite field in the North Qinling orogenic belt(NQB),China,hosts the most abundant LCT pegmatites.However,their emplacement conditions and structural control remain unexplored.In this contribution,we investigated it combining pegmatite orientation measurement with oxygen isotope geothermometry and fluid inclusion study.The orientations of type A1 pegmatites(P_(f)<σ_(2))are predominantly influenced by P-and T-fractures due to simple shearing in Shiziping dextral thrust shear zone during D_(2)deformation,whereas type A2 pegmatites(contemporaneous with D_(4))are governed by hydraulic fractures aligned with S_(0)and S_(0+1)stemming from fluid pressure(P_(f)<σ_(2)).Additionally,type B pegmatites(P_(f)≤σ_(2))exhibit orientations shaped by en echelon extensional fractures in local ductile shear zones(contemporaneous with D_(3)).The albite-quartz oxygen isotope geothermometry and microthermometric analysis of fluid inclusions in elbaites from the latest pegmatites(including types B and A2)suggest that the crystallization P-T for late magmatic and hydrothermal stages are 527.5-559.2℃,320℃,3.1-3.6 kbar and 2.0 kbar,respectively.Our observations along with previous studies suggest that the genesis of the LCT pegmatites was a long-term,multi-stage event during early Paleozoic orogeny(including the collision stage)of the NQB,and was facilitated by various local fractures.

Ambient-Condition Strategy for Production of Hollow Ga_(2)O_(3)@rGO Crystalline Nanostructures Toward Efficient Lithium Storage

Crystallineγ-Ga_(2)O_(3)@rGO core-shell nanostructures are synthesized in gram scale,which are accomplished by a facile sonochemical strategy under ambient condition.They are composed of uniformγ-Ga_(2)O_(3)nanospheres encapsulated by reduced graphene oxide(rGO)nanolayers,and their formation is mainly attributed to the existed opposite zeta potential between the Ga_(2)O_(3)and rGO.The as-constructed lithium-ion batteries(LIBs)based on as-fabricatedγ-Ga_(2)O_(3)@rGO nanostructures deliver an initial discharge capacity of 1000 mAh g^(-1)at 100 mA g^(-1)and reversible capacity of 600 mAh g^(-1)under 500 mA g^(-1)after 1000 cycles,respectively,which are remarkably higher than those of pristineγ-Ga_(2)O_(3)with a much reduced lifetime of 100 cycles and much lower capacity.Ex situ XRD and XPS analyses demonstrate that the reversible LIBs storage is dominant by a conversion reaction and alloying mechanism,where the discharged product of liquid metal Ga exhibits self-healing ability,thus preventing the destroy of electrodes.Additionally,the rGO shell could act robustly as conductive network of the electrode for significantly improved conductivity,endowing the efficient Li storage behaviors.This work might provide some insight on mass production of advanced electrode materials under mild condition for energy storage and conversion applications.

Autologous hematopoietic stem cell transplantation conditioning regimens and chimeric antigen receptor T cell therapy in various diseases

Conditioning regimens employed in autologous stem cell transplantation have been proven useful in various hematological disorders and underlying malignancies;however,despite being efficacious in various instances,negative consequences have also been recorded.Multiple conditioning regimens were extracted from various literature searches from databases like PubMed,Google scholar,EMBASE,and Cochrane.Conditioning regimens for each disease were compared by using various end points such as overall survival(OS),progression free survival(PFS),and leukemia free survival(LFS).Variables were presented on graphs and analyzed to conclude a more efficacious conditioning regimen.In multiple myeloma,the most effective regimen was high dose melphalan(MEL)given at a dose of 200/mg/m2.The comparative results of acute myeloid leukemia were presented and the regimens that proved to be at an admirable position were busulfan(BU)+MEL regarding OS and BU+VP16 regarding LFS.In case of acute lymphoblastic leukemia(ALL),BU,fludarabine,and etoposide(BuFluVP)conferred good disease control not only with a paramount improvement in survival rate but also low risk of recurrence.However,for ALL,chimeric antigen receptor(CAR)T cell therapy was preferred in the context of better OS and LFS.With respect to Hodgkin’s lymphoma,mitoxantrone(MITO)/MEL overtook carmustine,VP16,cytarabine,and MEL in view of PFS and vice versa regarding OS.Non-Hodgkin’s lymphoma patients were administered MITO(60 mg/m2)and MEL(180 mg/m2)which showed promising results.Lastly,amyloidosis was considered,and the regimen that proved to be competent was MEL 200(200 mg/m2).This review article demonstrates a comparison between various conditioning regimens employed in different diseases.

Attribute Reduction of Hybrid Decision Information Systems Based on Fuzzy Conditional Information Entropy

The presence of numerous uncertainties in hybrid decision information systems(HDISs)renders attribute reduction a formidable task.Currently available attribute reduction algorithms,including those based on Pawlak attribute importance,Skowron discernibility matrix,and information entropy,struggle to effectively manages multiple uncertainties simultaneously in HDISs like the precise measurement of disparities between nominal attribute values,and attributes with fuzzy boundaries and abnormal values.In order to address the aforementioned issues,this paper delves into the study of attribute reduction withinHDISs.First of all,a novel metric based on the decision attribute is introduced to solve the problem of accurately measuring the differences between nominal attribute values.The newly introduced distance metric has been christened the supervised distance that can effectively quantify the differences between the nominal attribute values.Then,based on the newly developed metric,a novel fuzzy relationship is defined from the perspective of“feedback on parity of attribute values to attribute sets”.This new fuzzy relationship serves as a valuable tool in addressing the challenges posed by abnormal attribute values.Furthermore,leveraging the newly introduced fuzzy relationship,the fuzzy conditional information entropy is defined as a solution to the challenges posed by fuzzy attributes.It effectively quantifies the uncertainty associated with fuzzy attribute values,thereby providing a robust framework for handling fuzzy information in hybrid information systems.Finally,an algorithm for attribute reduction utilizing the fuzzy conditional information entropy is presented.The experimental results on 12 datasets show that the average reduction rate of our algorithm reaches 84.04%,and the classification accuracy is improved by 3.91%compared to the original dataset,and by an average of 11.25%compared to the other 9 state-of-the-art reduction algorithms.The comprehensive analysis of these research results clearly indicates that our algorithm is highly effective in managing the intricate uncertainties inherent in hybrid data.

Spatio-temporal characteristics of genotoxicity in the Yangtze River under the background of COVID-19 pandemic

The global spread of viruses can lead to the release of large amounts of disinfectants or antiviral drugs into the water environment.The resulting disinfection byproducts(DBPs)and residual antiviral drugs,acting as genotoxic substances or their precursors,may pose risks to aquatic animals and drinking water sources;however,to date,no studies have analyzed the changes in genotoxicity in the Yangtze River before and after the epidemic.In the present study,water and sediment samples from the Yangtze River were collected during different seasons,just before and after the outbreak of COVID-19,and were assessed using the SOS/umu test(with and without liver S9).The results indicated that water samples exhibited more pronounced genotoxicity than did sediments,with direct genotoxicity being the primary factor.Additionally,there were significant regional differences,with notably greater genotoxicity observed in the upper Yangtze River than in the lower reaches before the COVID-19 epidemic.However,this trend was reversed six to ten months later,suggesting the accumulation of DBPs or antiviral drugs after the COVID-19 pandemic.Moreover,the risk quotient indicated that 65%of the water samples posed a high risk for Paramecium caudatum,whereas 71%of the samples posed a medium risk for Danio rerio,thereby representing a potential threat to the ecological security of the Yangtze River.In conclusion,this study,at the basin scale,revealed the impacts of COVID-19 on the Yangtze River,highlighting the need to prevent DBPs and pharmaceutical pollution during similar events in the future.

Improved spatio-temporal alignment measurement method for hull deformation

In this paper,an improved spatio-temporal alignment measurement method is presented to address the inertial matching measurement of hull deformation under the coexistence of time delay and large misalignment angle.Large misalignment angle and time delay often occur simultaneously and bring great challenges to the accurate measurement of hull deformation in space and time.The proposed method utilizes coarse alignment with large misalignment angle and time delay estimation of inertial measurement unit modeling to establish a brand-new spatiotemporal aligned hull deformation measurement model.In addition,two-step loop control is designed to ensure the accurate description of dynamic deformation angle and static deformation angle by the time-space alignment method of hull deformation.The experiments illustrate that the proposed method can effectively measure the hull deformation angle when time delay and large misalignment angle coexist.

Utilizing spatio-temporal feature fusion in videos for detecting the fluidity of coal water slurry

The fluidity of coal-water slurry(CWS)is crucial for various industrial applications such as long-distance transportation,gasification,and combustion.However,there is currently a lack of rapid and accurate detection methods for assessing CWS fluidity.This paper proposed a method for analyzing the fluidity using videos of CWS dripping processes.By integrating the temporal and spatial features of each frame in the video,a multi-cascade classifier for CWS fluidity is established.The classifier distinguishes between four levels(A,B,C,and D)based on the quality of fluidity.The preliminary classification of A and D is achieved through feature engineering and the XGBoost algorithm.Subsequently,convolutional neural networks(CNN)and long short-term memory(LSTM)are utilized to further differentiate between the B and C categories which are prone to confusion.Finally,through detailed comparative experiments,the paper demonstrates the step-by-step design process of the proposed method and the superiority of the final solution.The proposed method achieves an accuracy rate of over 90%in determining the fluidity of CWS,serving as a technical reference for future industrial applications.

Recent advances in spatio-temporally controllable systems for management of glioma

Malignant glioma remains one of the most aggressive intracranial tumors with devastating clinical outcomes despite the great advances in conventional treatment approaches,including surgery and chemotherapy.Spatio-temporally controllable approaches to glioma are now being actively investigated due to the preponderance,including spatio-temporal adjustability,minimally invasive,repetitive properties,etc.External stimuli can be readily controlled by adjusting the site and density of stimuli to exert the cytotoxic on glioma tissue and avoid undesired injury to normal tissues.It is worth noting that the removability of external stimuli allows for on-demand treatment,which effectively reduces the occurrence of side effects.In this review,we highlight recent advancements in drug delivery systems for spatio-temporally controllable treatments of glioma,focusing on the mechanisms and design principles of sensitizers utilized in these controllable therapies.Moreover,the potential challenges regarding spatio-temporally controllable therapy for glioma are also described,aiming to provide insights into future advancements in this field and their potential clinical applications.

IoT-based green-smart photovoltaic system under extreme climatic conditions for sustainable energy development

To realize carbon neutrality,there is an urgent need to develop sustainable,green energy systems(especially solar energy systems)owing to the environmental friendliness of solar energy,given the substantial greenhouse gas emissions from fossil fuel-based power sources.When it comes to the evolution of intelligent green energy systems,Internet of Things(IoT)-based green-smart photovoltaic(PV)systems have been brought into the spotlight owing to their cutting-edge sensing and data-processing technologies.This review is focused on three critical segments of IoT-based green-smart PV systems.First,the climatic parameters and sensing technologies for IoT-based PV systems under extreme weather conditions are presented.Second,the methods for processing data from smart sensors are discussed,in order to realize health monitoring of PV systems under extreme environmental conditions.Third,the smart materials applied to sensors and the insulation materials used in PV backsheets are susceptible to aging,and these materials and their aging phenomena are highlighted in this review.This review also offers new perspectives for optimizing the current international standards for green energy systems using big data from IoT-based smart sensors.

Epidemic Characteristics and Spatio-Temporal Patterns of HFRS in Qingdao City,China,2010-2022

Objective This study investigated the epidemic characteristics and spatio-temporal dynamics of hemorrhagic fever with renal syndrome(HFRS)in Qingdao City,China.Methods Information was collected on HFRS cases in Qingdao City from 2010 to 2022.Descriptive epidemiologic,seasonal decomposition,spatial autocorrelation,and spatio-temporal cluster analyses were performed.Results A total of 2,220 patients with HFRS were reported over the study period,with an average annual incidence of 1.89/100,000 and a case fatality rate of 2.52%.The male:female ratio was 2.8:1.75.3%of patients were aged between 16 and 60 years old,75.3%of patients were farmers,and 11.6%had both“three red”and“three pain”symptoms.The HFRS epidemic showed two-peak seasonality:the primary fall-winter peak and the minor spring peak.The HFRS epidemic presented highly spatially heterogeneous,street/township-level hot spots that were mostly distributed in Huangdao,Pingdu,and Jiaozhou.The spatio-temporal cluster analysis revealed three cluster areas in Qingdao City that were located in the south of Huangdao District during the fall-winter peak.Conclusion The distribution of HFRS in Qingdao exhibited periodic,seasonal,and regional characteristics,with high spatial clustering heterogeneity.The typical symptoms of“three red”and“three pain”in patients with HFRS were not obvious.

Dynamic adaptive spatio-temporal graph network for COVID-19 forecasting

Appropriately characterising the mixed space-time relations of the contagion process caused by hybrid space and time factors remains the primary challenge in COVID-19 forecasting.However,in previous deep learning models for epidemic forecasting,spatial and temporal variations are captured separately.A unified model is developed to cover all spatio-temporal relations.However,this measure is insufficient for modelling the complex spatio-temporal relations of infectious disease transmission.A dynamic adaptive spatio-temporal graph network(DASTGN)is proposed based on attention mechanisms to improve prediction accuracy.In DASTGN,complex spatio-temporal relations are depicted by adaptively fusing the mixed space-time effects and dynamic space-time dependency structure.This dual-scale model considers the time-specific,space-specific,and direct effects of the propagation process at the fine-grained level.Furthermore,the model characterises impacts from various space-time neighbour blocks under time-varying interventions at the coarse-grained level.The performance comparisons on the three COVID-19 datasets reveal that DASTGN achieves state-of-the-art results with a maximum improvement of 17.092%in the root mean-square error and 11.563%in the mean absolute error.Experimental results indicate that the mechanisms of designing DASTGN can effectively detect some spreading characteristics of COVID-19.The spatio-temporal weight matrices learned in each proposed module reveal diffusion patterns in various scenarios.In conclusion,DASTGN has successfully captured the dynamic spatio-temporal variations of COVID-19,and considering multiple dynamic space-time relationships is essential in epidemic forecasting.